WEAR LEVEL ESTIMATION IN ANALOG MEMORY CELLS

Abstract

A method for operating a memory includes applying at least one pulse to a group of analog memory cells, so as to cause the memory cells in the group to assume respective storage values. After applying the pulse, the respective storage values are read from the memory cells in the group. One or more statistical properties of the read storage values are computed. A wear level of the group of the memory cells is estimated responsively to the statistical properties.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of U.S. Provisional Patent Application 61/016,816, filed Dec. 27, 2007, U.S. Provisional Patent Application 61/026,211, filed Feb. 5, 2008, and U.S. Provisional Patent Application 61/078,780, filed Jul. 8, 2008, whose disclosures are incorporated herein by reference.

FIELD OF THE INVENTION

The present invention relates generally to memory devices, and particularly to methods and systems for estimating wear levels of analog memory cells.

BACKGROUND OF THE INVENTION

Several types of memory devices, such as Flash memories, use arrays of analog memory cells for storing data. Each analog memory cell stores a quantity of an analog value, also referred to as a storage value, such as an electrical charge or voltage. The storage value represents the information stored in the cell. In Flash memories, for example, each analog memory cell holds a certain amount of electrical charge. The range of possible analog values is typically divided into regions, each region corresponding to one or more data bit values. Data is written to an analog memory cell by writing a nominal analog value that corresponds to the desired bit or bits.

Some memory devices, which are commonly referred to as Single-Level Cell (SLC) devices, store a single bit of information in each memory cell, i.e., each memory cell can be programmed to assume two possible memory states. Higher-density devices, often referred to as Multi-Level Cell (MLC) devices, store two or more bits per memory cell, i.e., can be programmed to assume more than two possible memory states.

Flash memory devices are described, for example, by Bez et al., in “Introduction to Flash Memory,” Proceedings of the IEEE, volume 91, number 4, April, 2003, pages 489-502, which is incorporated herein by reference. Multi-level Flash cells and devices are described, for example, by Eitan et al., in “Multilevel Flash Cells and their Trade-Offs,” Proceedings of the 1996 IEEE International Electron Devices Meeting (IEDM), New York, N.Y., pages 169-172, which is incorporated herein by reference. The paper compares several kinds of multilevel Flash cells, such as common ground, DINOR, AND, NOR and NAND cells.

Eitan et al., describe another type of analog memory cell called Nitride Read Only Memory (NROM) in “Can NROM, a 2-bit, Trapping Storage NVM Cell, Give a Real Challenge to Floating Gate Cells?” Proceedings of the 1999 International Conference on Solid State Devices and Materials (SSDM), Tokyo, Japan, Sep. 21-24, 1999, pages 522-524, which is incorporated herein by reference. NROM cells are also described by Maayan et al., in “A 512 Mb NROM Flash Data Storage Memory with 8 MB/s Data Rate”, Proceedings of the 2002 IEEE International Solid-State Circuits Conference (ISSCC 2002), San Francisco, Calif., Feb. 3-7, 2002, pages 100-101, which is incorporated herein by reference. Other exemplary types of analog memory cells are Floating Gate (FG) cells, Ferroelectric RAM (FRAM) cells, magnetic RAM (MRAM) cells, Charge Trap Flash (CTF) and phase change RAM (PRAM, also referred to as Phase Change Memory—PCM) cells. FRAM, MRAM and PRAM cells are described, for example, by Kim and Koh in “Future Memory Technology including Emerging New Memories,” Proceedings of the 24^th International Conference on Microelectronics (MIEL), Nis, Serbia and Montenegro, May 16-19, 2004, volume 1, pages 377-384, which is incorporated herein by reference.

The performance of some types of analog memory cells deteriorates as a function of use (e.g., as a function of the number of Programming and Erasure (P/E) cycles applied to the cells). This property is sometimes referred to as cell wearing. Some data storage techniques attempt to distribute cell wearing evenly among different groups of memory cells. These techniques are commonly referred to as wear leveling techniques.

For example, U.S. Patent Application Publication 2007/0050536, whose disclosure is incorporated herein by reference, describes a system comprising a non-volatile memory, a controller and a wear leveling table. The non-volatile memory includes one or more memory blocks to store data. The controller allocates one or more of the memory blocks to store data. The wear-leveling table is populated with pointers to unallocated memory blocks in the non-volatile memory. The controller identifies one or more pointers in the wear-leveling table and allocates the unallocated memory blocks associated with the identified pointers for the storage of data.

U.S. Patent Application Publication 2006/0203546, whose disclosure is incorporated herein by reference, describes a method of achieving wear leveling in a Flash memory. For each block of the memory, a number is calculated that is a function of the number of times the block has been erased and of the number of times at least one other block has been erased. The numbers are stored in a memory device that includes the memory. The numbers are updated as needed when blocks are erased. Blocks are selected to be erased in accordance with their numbers.

PCT International Publication WO 2008/077284, whose disclosure is incorporated herein by reference, describes a wear leveling method for non-volatile memory. An embodiment includes counting erase cycles for each of a set of multiple memory blocks of a non-volatile memory, including incrementing a first count for a physical block address of the memory block. If the memory block is not a spare memory block, a second count for a logical block address of the memory block is incremented. The method also determines whether the memory has uneven wear of memory blocks based on the counting of the erase cycles of the plurality of memory blocks.

U.S. Patent Application Publication 2007/0208904, whose disclosure is incorporated herein by reference, describes a wear leveling apparatus, which uniformly distributes wear over a nonvolatile memory containing a plurality of memory blocks. The apparatus includes a memory unit for storing a record of cold block candidates in the nonvolatile memory and a control unit configured to update the memory unit and release the cold block candidates under a threshold condition. The control unit selects a new memory block to replace one cold block candidate in the memory unit when the cold block candidate is matched with a written address in a write command for the nonvolatile memory. The cold block candidates remaining in the memory unit are identified as cold blocks when the nonvolatile memory has been written more than a predetermined write count threshold. The memory blocks with infrequent erasure can be identified and released to uniformly distribute wear over the nonvolatile memory.

U.S. Pat. No. 6,230,233, whose disclosure is incorporated herein by reference, describes a mass storage system made of Flash memory cells organized into blocks, the blocks in turn being grouped into memory banks. The system is managed to even out the numbers of erase and rewrite cycles experienced by the memory banks in order to extend the service lifetime of the system. Relative use of the memory banks is monitored and, in response to detection of uneven use, memory banks have their physical addresses periodically swapped for each other in order to even out their use over the lifetime of the memory.

U.S. Pat. No. 7,441,067, whose disclosure is incorporated herein by reference, describes a re-programmable non-volatile memory system, such as a Flash system, having its memory cells grouped into blocks of cells that are simultaneously erasable. The system is operated in a manner to level out the wear of the individual blocks through repetitive erasing and re-programming. Wear leveling may be accomplished without use of counts of the number of times the individual blocks experience erase and re-programming but such counts can optionally aid in carrying out the wear leveling process. Individual active physical blocks are chosen to be exchanged with those of an erased block pool in a predefined order.

U.S. Pat. No. 7,453,737, whose disclosure is incorporated herein by reference, describes a non-volatile memory device and programming process. The programming process increases the programming voltage of successive programming cycles in relation to the percentage of the data bits that failed programming verification during the previous programming cycle and were not correctly programmed into the memory array.

SUMMARY OF THE INVENTION

An embodiment of the present invention provides a method for operating a memory, the method including:

applying at least one pulse to a group of analog memory cells so as to cause the memory cells in the group to assume respective storage values;

after applying the pulse, reading the respective storage values from the memory cells in the group;

computing one or more statistical properties of the read storage values; and

estimating a wear level of the group of the memory cells responsively to the statistical properties.

In some embodiments, estimating the wear level includes estimating a number of Programming and Erasure (P/E) cycles that were previously applied to the group of the memory cells. Estimating the wear level may include predefining a mapping between values of the statistical properties and respective wear levels, and determining the wear level that best matches the computed statistical properties according to the predefined mapping. In an embodiment, computing the statistical properties includes calculating at least one property selected from a group of properties consisting of a mean and a variance of the read storage values.

In a disclosed embodiment, the method includes storing data in the group of the analog memory cells by applying a sequence of Programming and Verification (P&V) operations to the memory cells in the group, applying the pulse includes applying a first programming operation in the sequence, and reading the storage values includes applying a first verification operation in the sequence. In another embodiment, estimating the wear level includes assessing multiple wear levels of multiple respective groups of the memory cells, and the method includes allocating at least one of the groups for storage of new data responsively to the assessed wear levels.

Additionally or alternatively, the method includes setting one or more parameters related to data storage in the group of the memory cells responsively to the estimated wear level. The parameters may include at least one parameter type selected from a group of types consisting of a read threshold, a verify threshold, an initial pulse amplitude of a sequence of programming pulses that are used for storing data in the memory cells, and an increment between successive programming pulses in the sequence of the programming pulses.

In some embodiments, the method includes setting a data storage density in the group of the memory cells responsively to the estimated wear level. Setting the data storage density may include selecting at least one property selected from a group of properties consisting of a number of programming levels and an Error Correction Code (ECC) used for data storage in the group.

In an embodiment, the method includes exchanging a message related to estimation of the wear level over an interface between the memory and a memory controller. Exchanging the message may include sending from the memory controller to the memory a command, which instructs the memory to apply the pulse. Additionally or alternatively, exchanging the message includes reporting the computed statistical properties from the memory to the memory controller. Further additionally or alternatively, exchanging the message may include reporting the estimated wear level from the memory to the memory controller.

In some embodiments, applying the at least one pulse includes applying one or more erasure pulses so as to cause at least some of the memory cells in the group to assume negative storage values.

Computing the statistical properties may include assessing a portion of the memory cells in the group that have assumed the negative storage values responsively to a given number of the erasure pulses. Additionally or alternatively, computing the statistical properties includes counting the erasure pulses needed to cause a given portion of the memory cells in the group to assume the negative storage values. Further additionally or alternatively, computing the statistical properties includes comparing the statistical properties before and after applying a given number of the erasure pulses.

In a disclosed embodiment, computing the statistical properties includes calculating the statistical properties for a subset of the memory cells in the group, which are initially programmed to a given programming level. In some embodiments, the memory cells in the group belong to a given word line, and reading the storage values includes causing at least some of the negative storage values to become positive by programming other memory cells, which belong to one or more word lines that neighbor the given word line.

There is additionally provided, in accordance with an embodiment of the present invention, a method for operating a memory, the method including:

predefining an action to be applied to one or more groups of analog memory cells in response to respective wear levels of the groups;

applying at least one pulse to a given group of the analog memory cells so as to cause the memory cells in the given group to assume respective storage values;

after applying the pulse, reading the respective storage values from the memory cells in the given group;

computing one or more statistical properties of the read storage values, which are indicative of a wear level of the memory cells in the given group; and

applying the predefined action responsively to the computed statistical properties.

There is further provided, in accordance with an embodiment of the present invention, apparatus, including:

a memory, which includes multiple analog memory cells; and

circuitry, which is coupled to apply at least one pulse to a group of the memory cells so as to cause the memory cells in the group to assume respective storage values, to read the respective storage values from the memory cells in the group after applying the pulse, to compute one or more statistical properties of the read storage values, and to estimate a wear level of the group of the memory cells responsively to the statistical properties.

There is further provided, in accordance with an embodiment of the present invention, apparatus, including:

a memory, which includes multiple analog memory cells; and

circuitry, which is coupled to hold a definition of an action to be applied to one or more groups of the memory cells in response to respective wear levels of the groups, to apply at least one pulse to a given group of the memory cells so as to cause the memory cells in the given group to assume respective storage values, to read the respective storage values from the memory cells in the given group after applying the pulse, to compute one or more statistical properties of the read storage values, which are indicative of a wear level of the memory cells in the given group, and to apply the action responsively to the computed statistical properties.

There is also provided, in accordance with an embodiment of the present invention, apparatus, including:

an interface, which is operative to communicate with a memory that includes multiple analog memory cells; and

circuitry, which is coupled to apply at least one pulse to a group of the memory cells so as to cause the memory cells in the group to assume respective storage values, to read the respective storage values from the memory cells in the group after applying the pulse, to compute one or more statistical properties of the read storage values, and to estimate a wear level of the group of the memory cells responsively to the statistical properties.

There is additionally provided, in accordance with an embodiment of the present invention apparatus, including:

an interface, which is operative to communicate with a memory that includes multiple analog memory cells; and

circuitry, which is coupled to hold a definition of an action to be applied to one or more groups of the memory cells in response to respective wear levels of the groups, to apply at least one pulse to a given group of the memory cells so as to cause the memory cells in the given group to assume respective storage values, to read the respective storage values from the memory cells in the given group after applying the pulse, to compute one or more statistical properties of the read storage values, which are indicative of a wear level of the memory cells in the given group, and to apply the action responsively to the computed statistical properties.

The present invention will be more fully understood from the following detailed description of the embodiments thereof, taken together with the drawings in which:

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram that schematically illustrates a memory system, in accordance with an embodiment of the present invention;

FIGS. 2A and 2B are graphs showing storage value distributions in groups of analog memory cells, in accordance with embodiments of the present invention; and

FIG. 3 is a flow chart that schematically illustrates a method for estimating a wear level of a group of analog memory cells, in accordance with an embodiment of the present invention.

DETAILED DESCRIPTION OF EMBODIMENTS

Overview

The performance of some types of analog memory cells degrades with use (e.g., as a function of the number of programming and erasure cycles applied to the cells). This phenomenon is commonly referred to as cell wearing. In some cases, the storage capacity of the cells deteriorates as a function of their wear level. At some point, a certain group of cells may become unusable due to extensive wearing. When operating a memory device that is partitioned into groups of memory cells, it is often desirable to assess the wear levels of the different cell groups.

Embodiments of the present invention that are described hereinbelow provide improved methods and systems for estimating the wear levels of groups of analog memory cells, and for taking various actions based on the estimated wear levels. The methods and systems described herein use the fact that analog memory cells tend to respond differently to programming or erasure pulses depending on their wear levels. In other words, when a given pulse is applied to a group of memory cells having a low wear level and to another cell group having a high wear level, the statistical distribution of storage values in the two groups will differ considerably.

In some embodiments that are described in detail below, the wear level of a given group of cells is estimated by applying one or more pulses to the memory cells in the group. The pulses cause the memory cells to assume respective storage values having a certain statistical distribution. The storage values of the memory cells are then read, and one or more statistical properties (e.g., mean and/or variance) of the read storage values are computed. The wear level of the group of memory cells is estimated based on the computed statistical properties.

In some embodiments, a memory device is partitioned into multiple cell groups, such as erasure blocks. The above-mentioned estimation process is repeated for different groups of memory cells, to produce respective wear level estimates of the different cell groups. The wear level estimates can be used, for example, for allocating low-wear cell groups to storage of new data. Additionally or alternatively, parameters of writing and reading operations in a given cell group can be set or modified based on the estimated wear level of the group. Several examples of parameters that can be modified as a function of wear level are described herein. This technique can improve the storage density of the memory device and the performance of read and write operations.

The methods and systems described herein provide fast and reliable wear level estimation. In particular, the disclosed techniques eliminate the need to track and store the history of each cell group (e.g., the number of programming and erasure cycles applied to each group of cells). In some cases, such historical information is unavailable, even when the memory has been used before extensively. In other cases, storing and tracking the historical information complicates the management of the memory and takes up storage space. Thus, the disclosed methods and systems simplify the wear level estimation process and eliminates the need to allocate storage space for storing such history tracking information.

System Description

FIG. 1 is a block diagram that schematically illustrates a memory system 20, in accordance with an embodiment of the present invention. System 20 can be used in various host systems and devices, such as in computing devices, cellular phones or other communication terminals, removable memory modules (“disk-on-key” devices), Solid State Disks (SSD), digital cameras, music and other media players and/or any other system or device in which data is stored and retrieved.

System 20 comprises a memory device 24, which stores data in a memory cell array 28. The memory cell array comprises multiple analog memory cells 32. In the context of the present patent application and in the claims, the term “analog memory cell” is used to describe any memory cell that holds a continuous, analog value of a physical parameter, such as an electrical voltage or charge. Array 28 may comprise analog memory cells of any kind, such as, for example, NAND, NOR and CTF Flash cells, PCM, NROM, FRAM, MRAM and DRAM cells. Memory cells 32 may comprise Single-Level Cells (SLC) or Multi-Level Cells (MLC, also referred to as multi-bit cells).

The charge levels stored in the cells and/or the analog voltages or currents written into and read out of the cells are referred to herein collectively as analog values or storage values. Although the embodiments described herein mainly address threshold voltages, the methods and systems described herein may be used with any other suitable kind of storage values.

System 20 stores data in the analog memory cells by programming the cells to assume respective memory states, which are also referred to as programming levels. The programming levels are selected from a finite set of possible levels, and each level corresponds to a certain nominal storage value. For example, a 2 bit/cell MLC can be programmed to assume one of four possible programming levels by writing one of four possible nominal storage values to the cell.

Memory device 24 comprises a reading/writing (R/W) unit 36, which converts data for storage in the memory device to storage values and writes them into memory cells 32. In alternative embodiments, the R/W unit does not perform the conversion, but is provided with voltage samples, i.e., with the storage values for storage in the cells. The R/W unit typically (although not necessarily) programs the cells using an iterative Program and Verify (P&V) process, as is known in the art. When reading data out of array 28, R/W unit 36 converts the storage values of memory cells 32 into digital samples having a resolution of one or more bits. Data is typically written to and read from the memory cells in groups that are referred to as pages. In some embodiments, the R/W unit can erase a group of cells 32 by applying one or more negative erasure pulses to the cells.

The storage and retrieval of data in and out of memory device 24 is performed by a Memory Signal Processor (MSP) 40. MSP 40 comprises an interface 44 for communicating with memory device 24, and a signal processing unit 48, which processes the data that is written into and read from device 24. In some embodiments, unit 48 produces the storage values for storing in the memory cells and provides these values to R/W unit 36. Alternatively, unit 48 provides the data for storage, and the conversion to storage values is carried out by the R/W unit internally to the memory device.

MSP 40 communicates with a host 52, for accepting data for storage in the memory device and for outputting data retrieved from the memory device. MSP 40, and in particular unit 48, may be implemented in hardware. Alternatively, MSP 40 may comprise a microprocessor that runs suitable software, or a combination of hardware and software elements.

The configuration of FIG. 1 is an exemplary system configuration, which is shown purely for the sake of conceptual clarity. Any other suitable memory system configuration can also be used. Elements that are not necessary for understanding the principles of the present invention, such as various interfaces, addressing circuits, timing and sequencing circuits and debugging circuits, have been omitted from the figure for clarity.

In the exemplary system configuration shown in FIG. 1, memory device 24 and MSP 40 are implemented as two separate Integrated Circuits (ICs). In alternative embodiments, however, the memory device and MSP may be integrated on separate semiconductor dies in a single Multi-Chip Package (MCP) or System on Chip (SoC). Further alternatively, some or all of the MSP circuitry may reside on the same die on which the memory array is disposed. Further alternatively, some or all of the functionality of MSP 40 can be implemented in software and carried out by a processor or other element of the host system, or by a suitable memory controller. In some implementations, a single MSP 40 may be connected to multiple memory devices 24. In yet another embodiment, some or all of the MSP functionality may be carried out by a separate unit, referred to as a memory extension, which acts as a slave of memory device 24.

Typically, the MSP (or any other controller or processor that carries out some or all of the methods described herein) comprises a general-purpose processor, which is programmed in software to carry out the functions described herein. The software may be downloaded to the processor in electronic form, over a network, for example, or it may, alternatively or additionally, be provided and/or stored on tangible media, such as magnetic, optical, or electronic memory.

Memory cells 32 of array 28 are typically arranged in a grid having multiple rows and columns, commonly referred to as word lines and bit lines, respectively. The array is typically divided into multiple pages, i.e., groups of memory cells that are programmed and read simultaneously. Cells are typically erased in groups of word lines that are referred to as erasure blocks.

In some embodiments, R/W unit 36 programs memory cells 32 using an iterative Program and Verify (P&V) process. In a typical P&V process, an entire memory page is written by applying a sequence of programming pulses to a group of memory cells that are mapped to this page. The amplitude of the programming pulses increases incrementally from pulse to pulse. The storage values programmed into the cells are read (“verified”) after each pulse, and the iterations continue until the desired levels are reached.

Wear Level Assessment Based on Storage Value Distribution

The performance of analog memory cells often deteriorates as a function of their wear level, e.g., as a function of the number of Programming and Erasure (P/E) cycles applied to the cells. For example, in Flash memory, charge traps are formed in boundary regions between the cells as the cells are being used. These charge traps accumulate with use and gradually degrade the storage capacity of the memory cells. In some cases, a group of memory cells (e.g., a block) may become unusable after a certain number of P/E cycles.

In the context of the present patent application and in the claims, the term “wear level” refers to a measure of the cumulative amount of use of a memory cell or group of memory cells, or to a measure of the change in the performance of the cell or cells due to such use. For example, the wear level of a group of cells can be quantified as the number of P/E cycles applied to the cells since the memory device was manufactured. Alternatively, wear level can be measured or quantified using any other suitable means. In some cases, the wear level of a group of cells can actually improve over time.

For example, in Flash memory, when no P/E operations are performed on the cells for a long period of time, charge tends to escape from the charge traps. As a result, the wear level of cells may improve if the cells are not accessed for a long time period. The methods and systems described herein enable differentiating between cell groups that have gone through a similar number of P/E cycles, but which have different wear levels because they differ in the time that elapsed since the last P/E operations.

Assessing the wear level of different cell groups is important for operating the memory device. The assessment can be used, for example, for storing data in the cells in a manner that matches their wear-dependent performance. As another example, wear level estimation can be used to determine when a certain cell group is unusable and should be taken out of service.

Embodiments of the present invention provide improved methods and systems for estimating the wear levels of groups of analog memory cells, and to take action upon the estimated wear levels. The disclosed methods and systems use the fact that analog memory cells typically respond differently to programming or erasure operations, depending on their wear levels. In some embodiments that are described herein, MSP 40 estimates the wear level of a given group of cells 32 in memory device 24 by instructing R/W unit 36 to apply one or more programming or erasure pulses to the cells in the group. The MSP then reads the storage values assumed by the cells as a result of these pulses, and computes one or more statistical properties (e.g., mean and/or variance) of the read storage values. The statistical properties are typically indicative of the wear level of the cell group. The description that follows refers mainly to wear level estimation using positive programming pulses. Alternative techniques for wear level estimation using negative erasure pulses are provided further below.

FIGS. 2A and 2B are graphs showing storage value distributions in groups of analog memory cells 32, in accordance with embodiments of the present invention. FIG. 2A shows the storage value distributions before and after applying a single programming pulse to a cell group having a relatively low wear level. FIG. 2B shows the distributions before and after applying a similar programming pulse to another cell group, which has a relatively high wear level. The two cell groups are assumed to be erased before applying the programming pulse. The characteristics of the programming pulse (e.g., its amplitude and duration) are assumed to be substantially the same for both cell groups.

In FIG. 2A, a distribution 60A represents the storage value distribution when the cells are at the erased level. In the present example, the erased cells have negative storage values that are distributed around a mean 64A. After applying the programming pulse, the cells in the low-wear group assume positive storage values, which are distributed according to a distribution 60B around a mean 64B.

In FIG. 2B, a distribution 60C shows the storage value distribution when the cells of the high-wear cell group are at the erased level, before applying the programming pulse. This distribution has a mean 64C. In the present example, although not necessarily, the storage value distribution of the erased cells differs from one group to the other. After the programming pulse is applied, the cells in the high-wear cell group assume positive storage values, which are distributed according to a distribution 60D around a mean 64D.

Comparing distribution 60B to distribution 60D demonstrates that the low-wear memory cells and the high-wear memory cells react differently to the programming pulse. As can be seen in FIGS. 2A and 2B, the statistical distribution of the storage values in the low-wear cell group after applying the programming pulse (distribution 60B) is distinctly different from the corresponding distribution in the high-wear cell group (distribution 60D). The storage value distribution in the high-wear cells typically has a larger variance and a higher mean, in comparison with the low-wear cell group. This sort of behavior is typical of Flash memory cells.

The dependence of the storage value distribution on wear level shown in FIGS. 2A and 2B is an exemplary dependence, which is chosen purely for the sake of conceptual clarity. The methods and systems described herein can be used with memory cells that respond differently to programming as a function of wear level in any other way.

FIG. 3 is a flow chart that schematically illustrates a method for estimating the wear level of a group of analog memory cells 32, e.g., of a given erasure block, in accordance with an embodiment of the present invention. The method begins with R/W unit 36 applying one or more open-loop programming pulses to the cells in the group, at a programming step 70. Typically, the R/W unit receives a command from MSP 40 to apply the pulses. In a typical implementation, the programming pulse applies to the word line in question a voltage in the range of 10-25V (e.g., 18V) for a duration of approximately 30 μS. Alternatively, however, any other suitable values can be used.

The term “open-loop” refers to the fact that no closed-loop verification is applied to the cells at this stage. Unlike Program and Verify (P&V) operation, in which pulses of different amplitudes are applied selectively to different cells in order to bring all cells to a target storage value, the purpose of step 70 is to deliberately differentiate between cells that react differently to programming pulses. Therefore, the R/W unit applies pulses having similar amplitudes to the different cells in the group. A technique that combines open-loop pulses with P&V programming of the cells is addressed further below.

The MSP reads the storage values from the memory cells in the group, at a reading step 74, and computes one or more statistical properties of the read storage values, at a statistics computation step 78. The MSP, using the R/W unit, typically performs one or more read operations, which compare the storage values of the cells to respective read thresholds. Based on the results of these read operations, the MSP estimates the statistical properties of the storage value distribution, such as the distribution mean and/or variance.

In some embodiments, the MSP performs rapid estimation by (1) applying a single programming pulse, (2) performing a single read operation, and (3) estimating the mean of the storage value distribution based on the results of the single read operation. The accuracy of estimating the statistical properties, however, typically depends on the number of read operations performed. Thus, in some embodiments, the MSP performs multiple read operations in order to assess the variance of the distribution and/or to estimate the mean more accurately.

Having computed the estimated statistical properties of the storage value distribution, the MSP estimates the wear level of the cell group as a function of the estimated statistical properties, at a wear level estimation step 82. In some embodiments, the MSP holds a predefined mapping between statistical property values and wear levels. For example, the predefined mapping may map mean and/or variance values of the storage values to respective numbers of P/E cycles. In these embodiments, the MSP determines the wear level value that best matches the estimated statistical properties of the cell group, in accordance with the predefined mapping. The mapping may be represented as a list, a look-up table, a function representation (e.g., a set of polynomial coefficients) or any other suitable representation.

As noted above, R/W unit 36 typically programs memory cells 32 using an iterative Program and Verify (P&V) process, in which the R/W unit applies to the cells a sequence of programming pulses that incrementally increase in amplitude. After each iteration, the R/W unit verifies the storage values assumed by the cells by reading the storage values and comparing them to one or more verify thresholds. At a given iteration, cells that have reached their target storage values are inhibited from receiving subsequent programming pulses, and programming continues for those cells that have not yet reached their intended storage values.

In some embodiments, the wear level estimation method of FIG. 3 can be combined as part of an iterative P&V process that stores data in the cells. For example, the MSP may use the programming pulse or pulses of the first P&V iteration as a sort of open-loop application of a programming pulse to the cells. Moreover, the MSP may use the first verification operation of the P&V process as a read operation, and compute the statistical properties of the storage value distribution based on the read results of the first P&V iteration. Moreover, the read results of subsequent P&V iterations can be used to further improve the accuracy of computing the statistical properties of the storage value distribution. For example, after a certain number of P&V iterations, the MSP may select a group of cells that received the same number of programming pulses and compute the statistical properties over the storage values of these cells.

The description of FIG. 3 above refers to a single group of cells. Typically, however, MSP 40 repeats the process of FIG. 3 in different cell groups, e.g., in different erasure blocks. As a result, the MSP may maintain a list of estimated wear levels of the different cell groups. The MSP can take various kinds of actions based on the estimated wear levels. For example, the MSP may give higher priority to low-wear blocks in storing data, in order to distribute cell wearing evenly among the different cell groups. The MSP may apply any suitable wear leveling technique, such as some of the techniques cited in the Background section above, for this purpose.

As another example, the MSP may define for each cell group one or more storage parameters, using which data is written to and/or read from the memory cells of the group. In some embodiments, the MSP sets the storage parameters of a given cell group based on the estimated wear level of the group. Using this technique, the writing and/or reading performance can be optimized in each cell group, according to the group's wear level.

Storage parameters that can be adapted based on wear level comprise, for example, values of read thresholds used in read operations and/or values of verify thresholds used in P&V programming operations. (Thus, the nominal storage values associated with the different programming levels, sometimes referred to as PV levels, may be adapted based on the wear level of the cells.) When data is stored in the cells using an iterative P&V process, the storage parameters may comprise various parameters of the P&V process, e.g., the initial amplitude of the programming pulses or the amplitude increment between successive programming pulses. Typically, low-wear cell groups can be programmed with rougher P&V parameters (e.g., a larger pulse increments or a smaller number of pulses), so as to improve programming speed. High-wear cells, on the other hand, should often be programmed more accurately in order to meet the desired data reliability. On the other hand, since high-wear cells are typically more responsive to programming pulses, a smaller number of pulses may often be sufficient for programming them.

As another example, the MSP may modify the storage density at which data is stored in the cells of a given group based on the estimated wear level of the group. Low-wear cells can often be programmed to store data at a higher density, and vice versa. The MSP can modify the storage density in a given cell group, for example, by modifying the number of programming levels or by modifying an Error Correction Code (ECC) that is used for encoding the stored data.

In some embodiments, the MSP holds a predefined mapping between statistical properties and storage parameters. This mapping may specify the estimated wear levels corresponding to the different statistical property and storage parameter values. In some embodiments, however, the wear levels may not be specified explicitly, such that the statistical properties are mapped to storage parameters directly.

In some embodiments, the MSP adaptively modifies the mapping of statistical properties to storage parameters during operation. For example, consider a scenario in which a certain statistical property value X (e.g., mean) is mapped to a certain estimated wear level, which in turn is translated to the use of a certain PV level. If the MSP concludes, during operation, that this PV level is marginal, it may adjust the mapping, so as to map statistical property value X to a different PV level value.

Wear Level Estimation Based on the Storage Value Distribution of Erased Cells

In the embodiments described above, MSP 40 estimates the wear level of a given cell group by applying positive pulses, i.e., pulses that increase the storage values of the cells. In alternative embodiments, however, the wear level can also be estimated by applying negative pulses, i.e., pulses that decrease the storage values of the cells. In particular, the wear level can be estimated by examining the statistical properties of the storage values of erased cells, before, during and/or after an erasure operation.

Typically, R/W unit 36 erases a group of memory cells 32 (e.g., an entire erasure block) by applying one or more erasure pulses having negative voltages to the word lines of these cells. These erasure pulses cause the cells to assume negative storage values (e.g., threshold voltages). See, for example, distribution 60A in FIG. 2A and distribution 60C in FIG. 2B. In some embodiments, the R/W unit erases cells using an iterative process, which somewhat resembles the P&V processes described above. In such an erasure process, the R/W unit applies a sequence of erasure pulses to the cells of a given erasure block. The erasure is typically verified per bit line. In other words, the R/W unit determines whether a particular bit line is positive, i.e., contains one or more cells whose storage value has not turned negative. Typically, erasure pulses are applied to the block until the number of positive bit lines drops below a specified number. If this target is not reached after a specified number of erase pulses, the erasure operation is regarded as failed.

As already noted above, the responsiveness of memory cells to positive programming pulses is typically indicative of the wear level of the cells. In many practical cases, similar behavior applies to negative erasure pulses, as well. In other words, the statistical distribution of the erased cells is often indicative of the wear level of the cells. Thus, in some embodiments, MSP 40 estimates the wear level of a group of erased cells based on one or more statistical properties of the storage values of the cells in the group. These statistical properties can be computed during and/or following the erasure operation.

Since in many memory devices it is difficult to measure negative storage values, the MSP may calculate the statistical properties (e.g., mean or variance) after applying a small number of erasure pulses, e.g., after the first erasure pulse. At that point, many of the cells still have positive storage values, which can be assessed using conventional read operations. In some embodiments, the MSP reads the storage values of the cells after each erasure iteration. In some embodiments, the MSP reads the cells before initiating the erasure operation, so that the relative effect of each erasure pulse can be estimated. (Note that different memory cells may be programmed to different levels prior to erasure.)

Additionally or alternatively, the MSP may estimate the wear level based on different measures regarding the storage values and the erasure operation, such as the number of cells whose storage values is still positive after a given number of erasure iterations, or the number of iterations needed to cause a given percentage of the cells (e.g., 99%) to assume negative storage values.

In some embodiments, the memory device provides an erase verification indication for each bit line, indicating whether or not the cells along the bit line are properly erased. The MSP may use these indications to estimate the number of cells whose storage value is positive. Alternatively, the MSP may perform a read operation with a certain (typically small) non-negative read threshold value. The read threshold value can be modified from one erasure iteration to another.

In some embodiments, the MSP carries out the following procedure for estimating the wear level of a given cell group:

• • The MSP selects a subset of the cells that are programmed to a certain programming level, and computes one or more statistical properties of the storage values of these cells (e.g., mean and/or variance) prior to erasure.
  • After each erasure iteration, the MSP reads the cells in the subset and re-computes the statistical properties. In some embodiments, the MSP can shift the storage value distribution to become more positive (i.e., to cause a larger portion of the distribution to occupy positive values) by modifying the bias voltages that are applied to non-read word lines in the block.
  • The MSP estimates the wear level of the cell group based on the different statistical properties before, during and/or after the erasure operation.

In some embodiments, the MSP may apply various techniques for estimating the storage value distribution of the cells after the erasure operation is completed. These embodiments typically involve assessing negative storage values. Several example techniques that can be used for this purpose are described in U.S. Patent Application Publication 2008/0181001, entitled “Memory Device with Negative Thresholds,” and in U.S. patent application Ser. No. 12/251,471, entitled “Compensation for Voltage Drifts in Analog Memory Cells.” Both of these patent applications are assigned to the assignee of the present patent application, and their disclosures are incorporated herein by reference.

For example, a negative storage value of a certain cell can sometimes be sensed by programming one or more neighboring cells to high storage values. The description that follows mainly addresses sensing of threshold voltages in NAND Flash cells, but this technique can also be used with other suitable memory cell types. The threshold voltages read from cells 32 are sometimes affected by the threshold voltages of other cells in the array. For example, the threshold voltage of a given cell may be affected by floating-gate coupling from neighboring cells, e.g., from horizontal neighbors on the same word line or from vertical neighbors on the same bit line. As another example, the memory cells in some NAND Flash devices are arranged in strings of cells that are connected in series, as is well-known in the art. The threshold voltage read from a given cell may be affected by the threshold voltages of other cells in the same NAND string.

In some embodiments, MSP 40 takes advantage of these effects in order to raise the negative threshold voltages of erased cells above zero, enabling the R/W unit to read these threshold voltages using conventional read operations. In some embodiments, the MSP programs one or more neighbors of a target erased cell to high values, in order to deliberately cause cross-coupling effects that would raise the threshold voltage of the target cell above 0V. For example, in order to sense the threshold voltages of the cells in a given word line, the MSP may program the cells in one or more neighboring word lines to high values (e.g., to the highest programming level supported by the memory device or to the maximum specified threshold voltage of the device). Typically, the MSP programs the neighboring word lines with high accuracy (e.g., with a high-resolution P&V process) in order to cause the threshold voltage distribution in the neighboring word lines to have a small variance.

Interface for Carrying Out Wear Level Estimation

In some embodiments, the MSP and the memory device support a command interface, which comprises commands that enable performing the wear level estimation methods described herein. For example, the command interface may comprise a command using which the MSP instructs the memory device to apply an open-loop pulse (or multiple open-loop pulses) to a specified group of cells.

In some embodiments, computation of the statistical properties of the storage value distribution is performed by circuitry internal to the memory device, and estimation of the wear level is performed by the MSP. In these embodiments, the command interface may comprise a command using which the memory device reports the computed statistical properties to the MSP. In alternative embodiments, the entire wear level estimation process is performed internally to the memory device. In these embodiments, the command interface may comprise a command using which the memory device reports the estimated wear level (e.g., the estimated number of P/E cycles) of a given cell group to the MSP. Alternatively, the memory device may report this value as part of a “program status” or an “erase status” message.

The command interface may be defined over the same physical interface (e.g., bus) that is used for data storage. Alternatively, the command interface may utilize a dedicated physical interface. Such a dedicated interface may be implemented in a number of ways. For example, the interface may comprise a bidirectional interface over which the MSP and memory device exchange commands and notifications with one another. Alternatively, the interface may comprise a unidirectional interface, which is directed from the memory device to the MSP. In these embodiments, the MSP can send commands to the memory device using the interface used for data storage, and receive notifications from the memory device over the dedicated unidirectional interface. In some embodiments, the unidirectional interface may comprise a logger, which logs the notifications and then outputs them continually or during predefined time periods, for example, during busy periods of the memory device. In addition to wear level-related information, the interface may be used for transmitting any other suitable kind of indications from the memory device to the MSP.

The dedicated command interface may be implemented using any suitable physical standard, such as UART, USB, FireWire. Typically, the interface utilizes a small number of pins, in order to reduce the overall pin count of the MSP and memory device. Further alternatively, the dedicated interface may comprise an analog signal interface. For example, the MSP may generate the open-loop pulses that are applied to the memory cells, and provide these pulses over an analog interface to the memory device.

In the embodiments described herein, the various wear level estimation functions are carried out by MSP 40 (or other controller that controls device 24). Alternatively, however, these functions can be carried out by elements internal to device 24 (e.g., by unit 36) and/or by elements external to the device (e.g., MSP 40). In the context of the present patent application and in the claims, these elements are referred to collectively as data storage circuitry, which is coupled to the memory and carries out the methods described herein. The disclosed methods can be partitioned among the different elements in any suitable way.

Although the embodiments described herein mainly address wear level estimation in solid state memory devices, the principles of the present invention can also be used for wear level estimation in other suitable storage media.

It will thus be appreciated that the embodiments described above are cited by way of example, and that the present invention is not limited to what has been particularly shown and described hereinabove. Rather, the scope of the present invention includes both combinations and sub-combinations of the various features described hereinabove, as well as variations and modifications thereof which would occur to persons skilled in the art upon reading the foregoing description and which are not disclosed in the prior art.

Claims

1. A method for operating a memory, the method comprising:

applying at least one pulse to a group of analog memory cells so as to cause the memory cells in the group to assume respective storage values;

after applying the pulse, reading the respective storage values from the memory cells in the group;

computing one or more statistical properties of the read storage values; and

estimating a wear level of the group of the memory cells responsively to the statistical properties.

2. The method according to claim 1, wherein estimating the wear level comprises estimating a number of Programming and Erasure (P/E) cycles that were previously applied to the group of the memory cells.

3. The method according to claim 1, wherein estimating the wear level comprises predefining a mapping between values of the statistical properties and respective wear levels, and determining the wear level that best matches the computed statistical properties according to the predefined mapping.

4. The method according to claim 1, wherein computing the statistical properties comprises calculating at least one property selected from a group of properties consisting of a mean and a variance of the read storage values.

5. The method according to claim 1, and comprising storing data in the group of the analog memory cells by applying a sequence of Programming and Verification (P&V) operations to the memory cells in the group, wherein applying the pulse comprises applying a first programming operation in the sequence, and wherein reading the storage values comprises applying a first verification operation in the sequence.

6. The method according to claim 1, wherein estimating the wear level comprises assessing multiple wear levels of multiple respective groups of the memory cells, and comprising allocating at least one of the groups for storage of new data responsively to the assessed wear levels.

7. The method according to claim 1, and comprising setting one or more parameters related to data storage in the group of the memory cells responsively to the estimated wear level.

8. The method according to claim 7, wherein the parameters comprise at least one parameter type selected from a group of types consisting of a read threshold, a verify threshold, an initial pulse amplitude of a sequence of programming pulses that are used for storing data in the memory cells, and an increment between successive programming pulses in the sequence of the programming pulses.

9. The method according to claim 1, and comprising setting a data storage density in the group of the memory cells responsively to the estimated wear level.

10. The method according to claim 9, wherein setting the data storage density comprises selecting at least one property selected from a group of properties consisting of a number of programming levels and an Error Correction Code (ECC) used for data storage in the group.

11. The method according to claim 1, and comprising exchanging a message related to estimation of the wear level over an interface between the memory and a memory controller.

12. The method according to claim 11, wherein exchanging the message comprises sending from the memory controller to the memory a command, which instructs the memory to apply the pulse.

13. The method according to claim 11, wherein exchanging the message comprises reporting the computed statistical properties from the memory to the memory controller.

14. The method according to claim 11, wherein exchanging the message comprises reporting the estimated wear level from the memory to the memory controller.

15. The method according to claim 1, wherein applying the at least one pulse comprises applying one or more erasure pulses so as to cause at least some of the memory cells in the group to assume negative storage values.

16. The method according to claim 15, wherein computing the statistical properties comprises assessing a portion of the memory cells in the group that have assumed the negative storage values responsively to a given number of the erasure pulses.

17. The method according to claim 15, wherein computing the statistical properties comprises counting the erasure pulses needed to cause a given portion of the memory cells in the group to assume the negative storage values.

18. The method according to claim 15, wherein computing the statistical properties comprises comparing the statistical properties before and after applying a given number of the erasure pulses.

19. The method according to claim 15, wherein computing the statistical properties comprises calculating the statistical properties for a subset of the memory cells in the group, which are initially programmed to a given programming level.

20. The method according to claim 15, wherein the memory cells in the group belong to a given word line, and wherein reading the storage values comprises causing at least some of the negative storage values to become positive by programming other memory cells, which belong to one or more word lines that neighbor the given word line.

21. A method for operating a memory, the method comprising:

predefining an action to be applied to one or more groups of analog memory cells in response to respective wear levels of the groups;

applying at least one pulse to a given group of the analog memory cells so as to cause the memory cells in the given group to assume respective storage values;

after applying the pulse, reading the respective storage values from the memory cells in the given group;

computing one or more statistical properties of the read storage values, which are indicative of a wear level of the memory cells in the given group; and

applying the predefined action responsively to the computed statistical properties.

22. Apparatus, comprising:

a memory, which comprises multiple analog memory cells; and

circuitry, which is coupled to apply at least one pulse to a group of the memory cells so as to cause the memory cells in the group to assume respective storage values, to read the respective storage values from the memory cells in the group after applying the pulse, to compute one or more statistical properties of the read storage values, and to estimate a wear level of the group of the memory cells responsively to the statistical properties.

23. The apparatus according to claim 22, wherein the circuitry is coupled to estimate the wear level by estimating a number of Programming and Erasure (P/E) cycles that were previously applied to the group of the memory cells.

24. The apparatus according to claim 22, wherein the circuitry is coupled to hold a predefined mapping between values of the statistical properties and respective wear levels, and to determine the wear level that best matches the computed statistical properties according to the predefined mapping.

25. The apparatus according to claim 22, wherein the circuitry is coupled to calculate at least one statistical property selected from a group of properties consisting of a mean and a variance of the read storage values.

26. The apparatus according to claim 22, wherein the circuitry is coupled to store data in the group of the analog memory cells by applying a sequence of Programming and Verification (P&V) operations to the memory cells in the group, to use a first programming operation in the sequence for applying the pulse, and to use a first verification operation in the sequence for reading the storage values.

27. The apparatus according to claim 22, wherein the circuitry is coupled to assess multiple wear levels of multiple respective groups of the memory cells, and to allocate at least one of the groups for storage of new data responsively to the assessed wear levels.

28. The apparatus according to claim 22, wherein the circuitry is coupled to set one or more parameters related to data storage in the group of the memory cells responsively to the estimated wear level.

29. The apparatus according to claim 28, wherein the parameters comprise at least one parameter type selected from a group of types consisting of a read threshold, a verify threshold, an initial pulse amplitude of a sequence of programming pulses that are used for storing data in the memory cells, and an increment between successive programming pulses in the sequence of the programming pulses.

30. The apparatus according to claim 22, wherein the circuitry is coupled to set a data storage density in the group of the memory cells responsively to the estimated wear level.

31. The apparatus according to claim 30, wherein the circuitry is coupled to set the data storage density by selecting at least one property selected from a group of properties consisting of a number of programming levels and an Error Correction Code (ECC) used for data storage in the group.

32. The apparatus according to claim 22, wherein the circuitry comprises a Read/Write (R/W) unit that is packaged in a memory device comprising the memory cells, and a memory controller that is separate from the memory device, and wherein the R/W unit and the memory controller are coupled to exchange a message related to estimation of the wear level.

33. The apparatus according to claim 32, wherein the memory controller is coupled to send to the R/W unit a command, which instructs the R/W unit to apply the pulse.

34. The apparatus according to claim 32, wherein the R/W unit is coupled to report the computed statistical properties to the memory controller.

35. The apparatus according to claim 32, wherein the R/W unit is coupled to report the estimated wear level to the memory controller.

36. The apparatus according to claim 22, wherein the at least one pulse comprises one or more erasure pulses, and wherein the circuitry is coupled to apply the erasure pulses so as to cause at least some of the memory cells in the group to assume negative storage values.

37. The apparatus according to claim 36, wherein the circuitry is coupled to assess a portion of the memory cells in the group that have assumed the negative storage values responsively to a given number of the erasure pulses.

38. The apparatus according to claim 36, wherein the circuitry is coupled to count the erasure pulses needed to cause a given portion of the memory cells in the group to assume the negative storage values.

39. The apparatus according to claim 36, wherein the circuitry is coupled to compare the statistical properties before and after applying a given number of the erasure pulses.

40. The apparatus according to claim 36, wherein the circuitry is coupled to calculate the statistical properties for a subset of the memory cells in the group, which are initially programmed to a given programming level.

41. The apparatus according to claim 36, wherein the memory cells in the group belong to a given word line, and wherein the circuitry is coupled to cause at least some of the negative storage values to become positive by programming other memory cells, which belong to one or more word lines that neighbor the given word line.

42. Apparatus, comprising:

a memory, which comprises multiple analog memory cells; and

circuitry, which is coupled to hold a definition of an action to be applied to one or more groups of the memory cells in response to respective wear levels of the groups, to apply at least one pulse to a given group of the memory cells so as to cause the memory cells in the given group to assume respective storage values, to read the respective storage values from the memory cells in the given group after applying the pulse, to compute one or more statistical properties of the read storage values, which are indicative of a wear level of the memory cells in the given group, and to apply the action responsively to the computed statistical properties.

43. Apparatus, comprising:

an interface, which is operative to communicate with a memory that includes multiple analog memory cells; and

circuitry, which is coupled to apply at least one pulse to a group of the memory cells so as to cause the memory cells in the group to assume respective storage values, to read the respective storage values from the memory cells in the group after applying the pulse, to compute one or more statistical properties of the read storage values, and to estimate a wear level of the group of the memory cells responsively to the statistical properties.

44. Apparatus, comprising:

an interface, which is operative to communicate with a memory that includes multiple analog memory cells; and

circuitry, which is coupled to hold a definition of an action to be applied to one or more groups of the memory cells in response to respective wear levels of the groups, to apply at least one pulse to a given group of the memory cells so as to cause the memory cells in the given group to assume respective storage values, to read the respective storage values from the memory cells in the given group after applying the pulse, to compute one or more statistical properties of the read storage values, which are indicative of a wear level of the memory cells in the given group, and to apply the action responsively to the computed statistical properties.